Emission reduction remains a critical objective in mitigating the environmental impact across various sectors, with the building sector contributing significantly to global emissions. While heat pump systems offer a sustainable alternative to traditional heating technologies, their operational efficiency is challenged by the varied nature of building infrastructures. This variability in infrastructure demands customized control strategies, as universal control approaches are often inefficient in addressing the unique characteristics and requirements of different buildings.
In today's digital world, Artificial Intelligence (AI) and Machine Learning (ML) are emerging as valuable tools, as they are capable of adapting to specific use cases, thus offering the ability to enhance operational efficiency. This presentation explores how AI and ML can address the inherent complexities and dynamic conditions of heat pump system operation. We will discuss two specific AI applications: a Reinforcement Learning-based self-optimizing defrost controller and a Data-driven Model Predictive Control for efficient and noise-reduced operation. These use cases effectively demonstrate how AI and ML can surpass conventional operational constraints, leading to an enhancement in operational efficiency.

Bio:
Jonas Klingebiel is a Research Associate at RWTH Aachen University's Institute for Energy Efficient Building and Indoor Climate.
With a master's degree in energy engineering from RWTH Aachen, he specializes in enhancing the efficiency of heat pumps through innovative control strategies. His research primarily focuses on the application of machine learning techniques, such as Reinforcement Learning and Data-Driven Model Predictive Control, to optimize heat pump operation. Additionally, Klingebiel is involved in the KI-Campus Fellowship Program, where he incorporates Machine Learning into higher education, specifically in the module Heat Pump System Engineering.

9TH INTERNATIONAL SYMPOSIUM

Jonas Klingebiel

Symposium keynote: Heat pump in the digital world: Exploiting the potential of AI and machine learning towards optimal operations

Link kopieret

Jonas Klingebiel